Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method, comprising: receiving, by a computing device, a request from a user for one or more recommendations; determining a location distribution from historical location information associated with the user, wherein the location distribution indicates one or more probability distributions for one or more locations; determining a time distribution from historical context information associated with the user, wherein the time distribution indicates one or more probability distributions for one or more contexts, wherein a respective context includes contextual variables corresponding to at least one of: a location; a time; a weather condition; and a venue type; determining a joint distribution of two or more of the contextual variables that are based on the location distribution and the time distribution, wherein the joint distribution comprises statistical attributes that characterize dependent relationships between the two or more of the contextual variables, the statistical attributes including one or more of: a mean value; a median value; a mode; a maximum value; and a minimum value of the contextual variables; determining a hypothetical context based on the statistical attributes of the joint distribution and a current context, wherein the hypothetical context indicates one or more user preferences that are outside of the user's current context and explicit historical contexts; producing one or more recommendations for the user based on the hypothetical context; constructing a mapping function that maps a user context to a hypothetical context, wherein constructing the mapping function involves using one or more of: historical contexts associated with the request; prior hypothetical contexts; and the user's interactions associated with the user's past requests for a recommendation; and providing the one or more recommendations to the user.
A system recommends items to a user by predicting a "hypothetical context" (e.g., future location, time, weather) beyond their current situation. It uses the user's historical location data to create a location probability distribution and historical context data (location, time, weather, venue type) to create a time probability distribution. A joint distribution combines these probabilities, revealing relationships between context variables (mean, median, etc.). The "hypothetical context" is determined based on this joint distribution and the user's current context, representing likely preferences beyond their immediate situation. Recommendations are generated using this predicted context. A mapping function learns to predict "hypothetical context" from a given context using historical data and user interactions, to improve the recommendation process.
2. The method of claim 1 , wherein determining the hypothetical context comprises: applying the mapping function to a current context; and producing the hypothetical context as an output of the mapping function.
Building upon the system that recommends items by predicting a "hypothetical context", the system determines the "hypothetical context" by applying a learned mapping function to the user's current context. The mapping function generates a predicted or expected context as output, which is then used to produce recommendations. This mapping function learns to predict the "hypothetical context" based on historical data and user interactions, allowing the system to anticipate changes in user context.
3. The method of claim 2 , wherein the construction of the mapping function is also based on a mapping relationship between the historical contexts and corresponding prior hypothetical contexts.
Expanding on the recommendation system's mapping function approach, the construction of the mapping function, which predicts future context for recommendations, considers the relationship between past contexts and their corresponding predicted "hypothetical contexts". In other words, the system learns not just from what happened, but also from previous predictions and if these predictions were accurate. The mapping function is built based on historical context and the corresponding "hypothetical context" pairs.
4. The method of claim 1 , wherein determining a respective hypothetical context comprises one or more of: determining a weather-condition distribution for the location distribution and the time distribution; determining a distribution over context-related preferences; and determining a distribution over conditions related to the hypothetical context that may be inferred from publicly-available data about the distribution over other contextual variables.
As part of determining the "hypothetical context" for recommendations, the system can determine the predicted "hypothetical context" by calculating either: a distribution of likely weather conditions for the location and time distributions; a distribution of the user's context-related preferences; or a distribution of conditions related to the "hypothetical context" that can be inferred from publicly available data about other context variables. The system uses predicted weather for likely locations and times, distribution of preferences or public data to make a better "hypothetical context" estimate.
5. The method of claim 4 , wherein determining the weather-condition distribution includes performing a lookup to determine a distribution over forecasted weather conditions corresponding to the location distribution and the time distribution.
Enhancing the weather prediction for "hypothetical context" determination, the system determines weather-condition distributions by looking up forecasted weather conditions based on the derived location and time probability distributions. This involves querying a weather service or data source to obtain a distribution of possible weather conditions for the predicted locations and times associated with the user. This allows the system to incorporate predicted weather into the recommendation process.
6. The method of claim 1 , wherein determining the time distribution comprises one or more of: examining one or more prior user requests associated with the location of the current context; determining the context associated with each prior user request; determining the hypothetical context associated with each prior user request; determining a mapping from the current context to the time distribution of the corresponding hypothetical context, wherein determining the mapping involves producing a time distribution based on the time associated with the current context and/or producing a time distribution based on the time associated with a calendar entry; constructing a mapping function based on the determined mapping; and applying the mapping function to the current context to determine the time distribution.
When determining the time distribution component of the "hypothetical context", the system performs these steps: examines prior requests associated with the current location, determines the context of those requests and corresponding "hypothetical contexts". It then maps the current context to a time distribution of the "hypothetical context", using the current time or calendar entries to predict future times. Finally, a mapping function is built based on this mapping to predict the time distribution from the current context.
7. The method of claim 1 , wherein estimating the location distribution comprises: collecting a set of candidate locations from location distributions that the user has previously specified and/or locations that the user has previously visited, excluding one or more of: candidate locations that have not been visited more than a predetermined number of times; location distributions that have not been specified more than a predetermined number of times; location distributions that were not specified during a time interval derived from the time associated with the current context; and candidate locations that were not visited during one or more time periods corresponding to the time distribution associated with the hypothetical context; and determining a distribution over locations from the set of candidate locations or one or more statistics of the set of candidate locations.
To estimate the location distribution when predicting a user's "hypothetical context", the system collects potential locations from the user's past specified locations and visited places. It then filters out locations visited infrequently, distributions rarely specified, distributions outside a relevant time interval, and locations not visited during periods corresponding to the predicted time distribution of the "hypothetical context". Finally, it determines a location distribution from the remaining candidate locations or calculates statistics from this set of locations.
8. The method of claim 7 , wherein the statistics of the set of candidate locations are determined in part from the type of venue associated with each candidate location; and wherein the location distribution includes locations not in the set of candidate locations but related to the type of venue of at least one candidate location.
Refining the location estimation process, the system determines statistics about potential locations partly based on the venue type associated with each location. The location distribution may also include locations *not* in the candidate set but related to the venue type of at least one candidate location. This allows the system to suggest related locations not previously visited by the user, but with similar venue types, improving recommendation diversity.
9. The method of claim 1 , wherein determining the hypothetical context involves: retrieving from a calendar a context whose associated time period substantially matches the time distribution of the hypothetical context; and setting other variables of the hypothetical context to the variables of the context retrieved from the calendar.
To determine the "hypothetical context", the system retrieves a context from the user's calendar whose time period closely matches the time distribution of the desired "hypothetical context". Then it sets the other variables of the "hypothetical context" (e.g. location, weather, venue type) to the values of the calendar entry's context. This effectively uses the user's calendar as a source of planned activities to guide recommendations.
10. The method of claim 1 , further comprising saving the hypothetical context to a calendar to facilitate subsequent hypothetical context determinations.
To improve future "hypothetical context" determinations, the system saves the determined "hypothetical context" to the user's calendar. This allows the system to learn from the chosen "hypothetical contexts" and refine its predictions over time, providing a feedback loop that enhances recommendation accuracy.
11. The method of claim 1 , further comprising: presenting the hypothetical context to the user; receiving a change to the hypothetical context from the user; and applying the change to the hypothetical context.
Before using the derived "hypothetical context" for recommendations, the system presents it to the user for approval. The user can then modify the "hypothetical context". These changes are applied to the "hypothetical context", thus the system takes user feedback to enhance accuracy.
12. A computer-readable non-transitory storage medium storing instructions that when executed by a computer cause the computer to perform a method, the method comprising: receiving a request from a user for one or more recommendations; determining a location distribution from historical location information associated with the user, wherein the location distribution indicates one or more probability distributions for one or more locations; determining a time distribution from historical context information associated with the user, wherein the time distribution indicates one or more probability distributions for one or more contexts, wherein a respective context includes contextual variables corresponding to at least one of: a location; a time; a weather condition; and a venue type; determining a joint distribution of two or more of the contextual variables that are based on the location distribution and the time distribution, wherein the joint distribution comprises statistical attributes that characterize dependent relationships between the two or more of the contextual variables, the statistical attributes including one or more of: a mean value; a median value; a mode; a maximum value; and a minimum value of the contextual variables; determining a hypothetical context based on the statistical attributes of the joint distribution and a current context, wherein the hypothetical context indicates one or more user preferences that are outside of the user's current context and explicit historical contexts; producing one or more recommendations for the user based on the hypothetical context; constructing a mapping function that maps a user context to a hypothetical context, wherein constructing the mapping function involves using one or more of: historical contexts associated with the request; prior hypothetical contexts; and the user's interactions associated with the user's past requests for a recommendation; and providing the one or more recommendations to the user.
A computer program stored on a non-transitory medium recommends items to a user by predicting a "hypothetical context" (e.g., future location, time, weather) beyond their current situation. It uses the user's historical location data to create a location probability distribution and historical context data (location, time, weather, venue type) to create a time probability distribution. A joint distribution combines these probabilities, revealing relationships between context variables (mean, median, etc.). The "hypothetical context" is determined based on this joint distribution and the user's current context, representing likely preferences beyond their immediate situation. Recommendations are generated using this predicted context. A mapping function learns to predict "hypothetical context" from a given context using historical data and user interactions, to improve the recommendation process.
13. The computer-readable non-transitory storage medium of claim 12 , wherein determining the hypothetical context comprises: applying the mapping function to a current context; and producing the hypothetical context as an output of the mapping function.
Building upon the computer program that recommends items by predicting a "hypothetical context", the system determines the "hypothetical context" by applying a learned mapping function to the user's current context. The mapping function generates a predicted or expected context as output, which is then used to produce recommendations. This mapping function learns to predict the "hypothetical context" based on historical data and user interactions, allowing the system to anticipate changes in user context.
14. The computer-readable non-transitory storage medium of claim 13 , wherein the construction of the mapping function is also based on a mapping relationship between the historical contexts and corresponding prior hypothetical contexts.
Expanding on the recommendation system's mapping function approach, the construction of the mapping function, which predicts future context for recommendations, considers the relationship between past contexts and their corresponding predicted "hypothetical contexts". In other words, the system learns not just from what happened, but also from previous predictions and if these predictions were accurate. The mapping function is built based on historical context and the corresponding "hypothetical context" pairs.
15. The computer-readable non-transitory storage medium of claim 12 , wherein determining a respective hypothetical context comprises one or more of: determining a weather-condition distribution for the location distribution and the time distribution; determining a distribution over context-related preferences; and determining a distribution over conditions related to the hypothetical context that may be inferred from publicly-available data about the distribution over other contextual variables.
As part of determining the "hypothetical context" for recommendations, the system can determine the predicted "hypothetical context" by calculating either: a distribution of likely weather conditions for the location and time distributions; a distribution of the user's context-related preferences; or a distribution of conditions related to the "hypothetical context" that can be inferred from publicly available data about other context variables. The system uses predicted weather for likely locations and times, distribution of preferences or public data to make a better "hypothetical context" estimate.
16. The computer-readable non-transitory storage medium of claim 15 , wherein determining the weather-condition distribution includes performing a lookup to determine a distribution over forecasted weather conditions corresponding to the location distribution and the time distribution.
Enhancing the weather prediction for "hypothetical context" determination, the system determines weather-condition distributions by looking up forecasted weather conditions based on the derived location and time probability distributions. This involves querying a weather service or data source to obtain a distribution of possible weather conditions for the predicted locations and times associated with the user. This allows the system to incorporate predicted weather into the recommendation process.
17. The computer-readable non-transitory storage medium of claim 12 , wherein determining the time distribution comprises one or more of: examining one or more prior user requests associated with the location of the current context; determining the context associated with each prior user request; determining the hypothetical context associated with each prior user request; determining a mapping from the current context to the time distribution of the corresponding hypothetical context, wherein determining the mapping involves producing a time distribution based on the time associated with the current context and/or producing a time distribution based on the time associated with a calendar entry; constructing a mapping function based on the determined mapping; and applying the mapping function to the current context to determine the time distribution.
When determining the time distribution component of the "hypothetical context", the system performs these steps: examines prior requests associated with the current location, determines the context of those requests and corresponding "hypothetical contexts". It then maps the current context to a time distribution of the "hypothetical context", using the current time or calendar entries to predict future times. Finally, a mapping function is built based on this mapping to predict the time distribution from the current context.
18. The computer-readable non-transitory storage medium of claim 12 , wherein estimating the location distribution comprises: collecting a set of candidate locations from location distributions that the user has previously specified and/or locations that the user has previously visited, excluding one or more of: candidate locations that have not been visited more than a predetermined number of times; location distributions that have not been specified more than a predetermined number of times; location distributions that were not specified during a time interval derived from the time associated with the current context; and candidate locations that were not visited during one or more time periods corresponding to the time distribution associated with the hypothetical context; and determining a distribution over locations from the set of candidate locations or one or more statistics of the set of candidate locations.
To estimate the location distribution when predicting a user's "hypothetical context", the system collects potential locations from the user's past specified locations and visited places. It then filters out locations visited infrequently, distributions rarely specified, distributions outside a relevant time interval, and locations not visited during periods corresponding to the predicted time distribution of the "hypothetical context". Finally, it determines a location distribution from the remaining candidate locations or calculates statistics from this set of locations.
19. The computer-readable non-transitory storage medium of claim 18 , wherein the statistics of the set of candidate locations are determined in part from the type of venue associated with each candidate location; and wherein the location distribution includes locations not in the set of candidate locations but related to the type of venue of at least one candidate location.
Refining the location estimation process, the system determines statistics about potential locations partly based on the venue type associated with each location. The location distribution may also include locations *not* in the candidate set but related to the venue type of at least one candidate location. This allows the system to suggest related locations not previously visited by the user, but with similar venue types, improving recommendation diversity.
20. The computer-readable non-transitory storage medium of claim 12 , wherein determining the hypothetical context involves: retrieving from a calendar a context whose associated time period substantially matches the time distribution of the hypothetical context; and setting other variables of the hypothetical context to the variables of the context retrieved from the calendar.
To determine the "hypothetical context", the system retrieves a context from the user's calendar whose time period closely matches the time distribution of the desired "hypothetical context". Then it sets the other variables of the "hypothetical context" (e.g. location, weather, venue type) to the values of the calendar entry's context. This effectively uses the user's calendar as a source of planned activities to guide recommendations.
21. The computer-readable non-transitory storage medium of claim 12 , wherein the method further comprises saving the hypothetical context to a calendar to facilitate subsequent hypothetical context determinations.
To improve future "hypothetical context" determinations, the system saves the determined "hypothetical context" to the user's calendar. This allows the system to learn from the chosen "hypothetical contexts" and refine its predictions over time, providing a feedback loop that enhances recommendation accuracy.
22. The computer-readable non-transitory storage medium of claim 12 , wherein the method further comprises: presenting the hypothetical context that is obtained from estimations and determinations of time, location and weather-condition distributions, to the user; receiving a change to the hypothetical context from the user; and applying the change to the hypothetical context.
Before using the derived "hypothetical context" that is obtained from estimations and determinations of time, location and weather-condition distributions, for recommendations, the system presents it to the user for approval. The user can then modify the "hypothetical context". These changes are applied to the "hypothetical context", thus the system takes user feedback to enhance accuracy.
23. An apparatus, comprising: a processor; a memory; a request receiving mechanism configured to receive a request from a user for one or more recommendations; a location estimation mechanism configured to determine a location distribution from historical location information associated with the user, wherein the location distribution indicates one or more probability distributions for one or more locations; a time estimation mechanism configured to determine a time distribution from historical context information associated with the user, wherein the time distribution indicates one or more probability distributions for one or more contexts, wherein a respective context includes contextual variables corresponding to at least one of: a location; a time; a weather condition; and a venue type; a joint distribution estimation mechanism configured to determine a joint distribution of two or more of the contextual variables that are based on the location distribution and the time distribution, wherein the joint distribution comprises statistical attributes that characterize dependent relationships between the two or more of the contextual variables, the statistical attributes including one or more of: a mean value; a median value; a mode; a maximum value; and a minimum value of the contextual variables; a hypothetical context estimation mechanism configured to determine a hypothetical context based on the statistical attributes of the joint distribution and a current context, wherein the hypothetical context indicates one or more user preferences that are outside of the user's current context and explicit historical contexts; a learning mechanism configured to construct a mapping function that maps a user context to a hypothetical context, wherein constructing the mapping function involves using one or more of: historical contexts associated with the request; prior hypothetical contexts; and the user's interactions associated with the user's past requests for a recommendation; and a recommendation mechanism configured to produce one or more recommendations for the user based on the hypothetical context, and to provide the one or more recommendations to the user.
The apparatus includes a processor and memory to receive a request for recommendations from a user. It estimates location probability distributions from the user's historical location data and time probability distributions from historical context data (location, time, weather, venue type). It uses these distributions to determine a joint distribution and estimate the "hypothetical context" based on that and current context. A learning mechanism creates a mapping function to predict the "hypothetical context". A recommendation mechanism produces recommendations using the "hypothetical context" and provides them to the user.
Unknown
October 28, 2014
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.